!! "moe" - routed inference between 3 different models without any tying

experimental MoE with 3 experts totalling 480m~ params router is roughly 70M params

no loss chart for this router trained on 15 samples

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support